About
About Me
- YoE: 5+
- Position: L6
- LC Solved: >800 problems (E:46%, M:45%, H:9%)
- Coupang problem in last 6 months = 8/15
About Job
- Position: L6
- Base: South Korea
- How did the interview take place? Virtually (Zoom + Hackerank)
- How did I get the interview? HR reached out to me via LinkedIn.
- They're willing to help me with the reallocation process (only apply for >= L6)
The Interview
1. Phone Screen (1h)
Usually, it is coding, but this one is not. It's simply chit-chat with their manager about my work experiences, projects, and how I handle things.
Reflection: Very chill, no pressure, and everything went well.
Response: 1 day
Feedback: N/A
2. Full Loop (1h each)
Note: Some of the interviews have interpretation (ENG-KOR), so there will be 3 people in the room, and some of them are ex-Meta / ex-Google.
2.1 System / Product Design
-
Design one of my projects from resume
I didn't expect this; however, I was heavily involved with this project, so I can answer it. -
Design Search Ranking
Same thing; I didn't expect this and never really studied about it. I was expect something like, design WhatsApp, Spotify, etc. However, I have some knowledge from the other team, which is the search and recommendation team, and how they communicate with other dependencies such as orders, ads, PDP, etc., which should be sufficient to give him the HLD.
2 designs in an hour?
I understand this is weird; however, the interviewer don't really ask you to do a deep dive, e.g., why use this database compared to others, what kind of data, etc. Therefore, I can finish it quickly.
2.2 Coding Design
Coding design is literally coding API feature. You don't need to compile, and feel free to use pseudocode. The foundation is the same; don't forget to ask for clarification, talk while coding, and present some corner cases.
Movie Ticket
Requirements:
- Don't store the data in DB, but store it in memory. You only have 1 machine.
- Code GET Search and POST Book Ticket.
- How do you handle race conditions?
I just treat this as Coding Algo with tag AdHoc, no problem. I managed to finish on time with pseudocode.
2.3 Coding Algo
Coding Algo is same what you've been practiced in LeetCode.
- Split Strings (Easy)
Similar to this https://leetcode.com/problems/split-strings-by-separator/ but simpler. The interviewer also asked me to compile the code. I can solved this under 1m.
Desc:
Get word(s) from a given string after you split it by space. The input chars can be anything.
input: "Aa12$. gg b c"
output: ["Aa12$.", "gg", "b", "c"]
input: "x"
output: ["x"]
input: ""
output: []- LFU Cache (Hard)
Exactly the same as this https://leetcode.com/problems/lfu-cache/description/, however, I only have like 20 minutes left to code after he explains too much about what the problem is, along with some unnecessary debate with my implementation where it actually doesn't change any complexity at all (my proposed solution is already optimal since I already AC it before). Thankfully, I managed to solve the code with a little bug (no compile).
2.4 Culture Fit
Talk about work experience and behavior. It's almost the same as a phone screen, I would say. However, there's a little miscommunication about what the interviewer asked, and I answered because I thought it was enough detail, but it turns out he demands more depth. Overall should be fine.
Reflection:
- Honestly, I almost YOLO everything because the product design and code design aren't what I expect, even though the HR already gave me the guideline (verbal), but it wasn't accurate enough. The code algo is evil; if you don't know how to LFU and are unable to solve it in less than 20 minutes, you're likely done. For culture, it is chill.
- I'm not comfortable with interpretation; I'm afraid they misinterpret something (worst case)
Response: 1 week
Feedback: Strong in tech, with minor improvement in leadership.
3. Leadership
This is a full behavioral question. The interviewer asked me a lot of questions, some of which were the same as the previous one. I don't have any difficulty explaining to the interviewer.
Reflection:
- I think this round is redundant since there's already a behavioral question on the phone screen and a full loop.
- This one is also done with interpretation, so yeah, you know.
Response: 2 weeks (w/ follow-up email)
Feedback: N/A
Result
Rejected, no feedback.
Either something went wrong with the last round or someone just better.